Wyniki wyszukiwania - BazTech

1

An adaptive island model of population for neuroevolutionary ship handling

Łącki Mirosław

Polish Maritime Research

|

2021

|

nr 4

142--150

EN

This study presents a method for the dynamic value assignment of evolutionary parameters to accelerate, automate and generalise the neuroevolutionary method of ship handling for different navigational tasks and in different environmental conditions. The island model of population is used in the modified neuroevolutionary method to achieve this goal. Three different navigational situations are considered in the simulation, namely, passing through restricted waters, crossing with another vessel and overtaking in the open sea. The results of the simulation examples show that the island model performs better than a single non-divided population and may accelerate some complex and dynamic navigational tasks. This adaptive island-based neuroevolutionary system used for the COLREG manoeuvres and for the finding safe ship’s route to a given destination in restricted waters increases the accuracy and flexibility of the simulation process. The time statistics show that the time of simulation of island NEAT was shortened by 6.8% to 27.1% in comparison to modified NEAT method.

2

Ship course-keeping with neuroevolutionary algorithms

Łącki M.

Zeszyty Naukowe Akademii Morskiej w Szczecinie

|

2018

|

nr 54 (126)

70--74

EN

The goal of research presented in this article is to check if a neuroevolutionary method with direct encoding is able to be a part of autopilot of the vessel. One of the important tasks of vessel autopilots is to keep a course as straight as possible or to bring the ship back on the route as efficiently as possible. In this paper, the adaptive neuroevolutionary autopilot is described and tested on a simulation model of a ferry. Neuroevolution is a combination of two different but related fields of artificial machine learning: evolution and neural networks. The combined method is very flexible and can be applied to other ship control tasks. The results of computer simulation of the neuroevolutionary course-keeping system have been included.

3

Intelligent Prediction of Ship Maneuvering

Łącki M.

TransNav : International Journal on Marine Navigation and Safety of Sea Transportation

|

2016

|

Vol. 10, no. 3

511--516

EN

In this paper the author presents an idea of the intelligent ship maneuvering prediction system with the usage of neuroevolution. This may be also be seen as the ship handling system that simulates a learning process of an autonomous control unit, created with artificial neural network. The control unit observes input signals and calculates the values of required parameters of the vessel maneuvering in confined waters. In neuroevolution such units are treated as individuals in population of artificial neural networks, which through environmental sensing and evolutionary algorithms learn to perform given task efficiently. The main task of the system is to learn continuously and predict the values of a navigational parameters of the vessel after certain amount of time, regarding an influence of its environment. The result of a prediction may occur as a warning to navigator to aware him about incoming threat.

4

Evolving co-adapted subcomponents in assembler encoding

Praczyk T.

International Journal of Applied Mathematics and Computer Science

|

2007

|

Vol. 17, no 4

549-563

EN

The paper presents a new Artificial Neural Network (ANN) encoding method called Assembler Encoding (AE). It assumes that the ANN is encoded in the form of a program (Assembler Encoding Program, AEP) of a linear organization and of a structure similar to the structure of a simple assembler program. The task of the AEP is to create a Connectivity Matrix (CM) which can be transformed into the ANN of any architecture. To create AEPs, and in consequence ANNs, genetic algorithms (GAs) are used. In addition to the outline of AE, the paper also presents a new AEP encoding method, i.e., the method used to represent the AEP in the form of a chromosome or a set of chromosomes. The proposed method assumes the evolution of individual components of AEPs, i.e., operations and data, in separate populations. To test the method, experiments in two areas were carried out, i.e., in optimization and in a predator-prey problem. In the first case, the task of AE was to create matrices which constituted a solution to the optimization problem. In the second case, AE was responsible for constructing neural controllers used to control artificial predators whose task was to capture a fast-moving prey.